The practice of compression has never been driven only by information theory. At least at an elementary level, information theory suggests techniques that require complete statistical knowledge of a class of data and even with this knowledge are computationally infeasible. Compression has benefited, for example, from efficient signal representations from the field of harmonic analysis, in particular advances in nonlinear approximation; computational techniques from signal processing; advanced techniques in statistical modeling and statistical inference; and insights and innovations from hands-on engineering.
The aim of this workshop is to continue and to accelerate the exchange of ideas between the various groups that contribute to compression.
This will include researchers in the following overlapping fields:
- source coding theory (information theorists)
- compression practice (signal processors)
- statistical modeling and inference (statisticians)
- harmonic analysis (applied mathematicians)
Talks that make progress in one field accessible to the spectrum of participants and those that are speculative are particularly encouraged.